Yes, a better error message sure would be nice. Surprising, Rexx is usually
better about such things!


ISTR that even PL/I gives a coherent message if you hit its maximum.
(384-yes, that's too many, although we hit it in a pathological case, where
we were passing tuples of three or four things each. So while it's too many
for sanity, it was useful as an illustration; we were setting our own
implementation limit of 100 tuples, as "way more than you should ever

For IBM-MAIN subscribe / signoff / archive access instructions,
send email to with the message: INFO IBM-MAIN

Reply via email to